Stimulation of the P2X7 receptor by ATP induces cell membrane depolarization, increase in intracellular Ca2+ concentration, and, in most cases, permeabilization of the cell membrane to molecules up to 900 Da. After the activation of P2X7, at least two phenomena occur: the opening of low-conductance (8 pS) cationic channels and pore formation. At least two conflicting hypotheses have been postulated to reconcile these findings: 1) the P2X7 pore is formed as a result of gradual permeability increase (dilation) of cationic channels, and 2) the P2X7 pore represents a distinct channel, possibly activated by a second messenger and not directly by extracellular nucleotides. In this study, we investigated whether second messengers are necessary to open the pore associated with the P2X7 receptor in cells that expressed the pore activity by using the patch-clamp technique in whole cell and cell-attached configurations in conjunction with fluorescent imaging. In peritoneal macrophages and 2BH4 cells, we detected permeabilization and single-channel currents in the cell-attached configuration when ATP was applied outside the membrane patch in a condition in which oxidized ATP and Lucifer yellow were maintained within the pipette. Our data support Ca2+ as a second messenger associated with pore formation because the permeabilization depended on the presence of intracellular Ca2+ and was blocked by BAPTA-AM. In addition, MAPK inhibitors (SB-203580 and PD-98059) blocked the permeabilization and single-channel currents in these cells. Together our data indicate that the P2X7 pore depends on second messengers such as Ca2+ and MAP kinases.
- pore formation
the first description of a physiological effect of extracellular ATP in the immune system was histamine release from mast cells (30). Later, it was shown that ATP-induced histamine release is concentration dependent (13) and that higher concentrations of ATP lead to permeabilization to progressively larger solutes across the cell membrane. This observation led the same authors to propose that channels opened by ATP could progressively increase their size similarly to the channels formed by certain antibiotics (14).
The receptor implicated in that phenomenon was later identified as the P2Z receptor on the basis of its unique properties: 1) sensitivity to extracellular ATP (ATP4−) and to its analog 2′,3′-O-(benzoyl-4-benzoyl)-ATP (BzATP), 2) association with channels permeable to small cations, 3) delayed plasma membrane permeabilization to larger molecules such as ethidium or propidium dye, and 4) induction of membrane blebbing and lysis in some cells (18, 25, 37, 42, 48, 52, 55).
Several groups of researchers sustained the notion that the permeabilizing receptor was a P2 receptor subclass distinct from the P2X receptors, until 1996, when Surprenant et al. (48) cloned a cDNA from rat brain that encoded a protein containing 595 amino acids. This protein showed strong homology with the P2X receptors but presented pharmacological and functional properties similar to those of the P2Z receptor and was named P2X7. When heterologously expressed, this cDNA produces an ionotropic receptor that, upon prolonged exposure to ATP, renders the plasma membrane permeable to the fluorescent dye YO-PRO-1 (630 Da).
However, some groups have had difficulty in detecting the low-selectivity pore in cells expressing the heterologous P2X7 receptor, suggesting that the pore-forming molecule might be a distinct entity (3, 14, 31a, 41a). Similarly, some cell types present a native P2 receptor that shares most pharmacological features of the P2X7 receptor but fails to produce permeabilization (33, 51). In keeping with this idea, Coutinho-Silva and Persechini (17) described a pore with a unitary conductance of 409 pS that had a pharmacological profile similar to that of the P2X7 pore. In that study, pore formation was detected only in the cell-attached configuration, which maintains the intracellular milieu relatively intact. Thus those authors suggested that this pore might not be associated directly with the P2Z/P2X7 receptor but instead might be activated by a second messenger (12, 17).
The mechanism underlying P2X7 pore formation is unknown, but there are at least two hypotheses. In one view, the channel formed by the P2X7 receptor, when activated, gradually dilates, increasing its permeability from small molecules to larger ones (up to 900 Da) (52, 53). The other hypothesis is that the channel formed by the activated P2X7 receptor induces the production or release of a second messenger that may autoactivate the P2X7 receptor or an independent pore-forming membrane protein (17, 41).
Using electrophysiological techniques in conjunction with fluorescent imaging, we have investigated whether such pore formation depends on second messengers. We also suggest the possible second messengers involved.
MATERIALS AND METHODS
The thymic epithelial cell line 2BH4, derived from C57B1/6 mouse thymus, was kindly provided by Dr. J. G. Amarante-Mendes (Dept. of Immunology, University of São Paulo, São Paulo, Brazil) (5). These cells were cultured in RPMI 1640 medium containing 10% fetal calf serum, penicillin (100 U/ml), and streptomycin (100 μg/ml) and plated on 35-mm petri dishes for 2–5 days until use. Thioglycolate-elicited macrophages obtained from the intraperitoneal cavity of Swiss mice were prepared in our laboratory. These cells were cultured in RPMI 1640 medium containing 10% fetal calf serum, penicillin (100 U/ml), and streptomycin (100 μg/ml) and plated on 35-mm petri dishes for 2–7 days until use. Cells were incubated at 37°C in a 5% CO2 humidified atmosphere.
Cell-attached and whole cell patch clamping was performed at 37°C using an Axopatch-1D amplifier (Axon Instruments, San Mateo, CA). Cells were transferred to a chamber mounted on a microscope stage. Patch pipettes (with 1.2-mm filament) were pulled from IBBL borosilicate glass capillaries (World Precision Instruments, New Haven, CT).
After a high-resistance (1–10 GΩ) seal was established by gentle suction, the cell membrane beneath the tip of the electrode was disrupted using additional suction. Currents obtained in the presence of agonist were not corrected for leakage, because this was negligible (currents in the absence of agonist were <0.1% of maximal agonist-induced currents). Recordings were accepted if the current and membrane conductance returned to within 1–5% of control values after agonist application, thus indicating that the large conductance increase was not due to cell lysis with loss of seal.
Series resistance was 6–10 MΩ for all experiments, and no compensation was applied for currents <400 pA. Above this level, currents were 85% compensated. Experiments in which the series resistance increased substantially during measurement were discarded. Cell capacitance (19.4 ± 2.06 pF; n = 82) was measured by applying a 20-mV hyperpolarizing pulse from a holding potential of −20 mV. Capacitive transient was then integrated and divided by the amplitude of the voltage step (20 mV).
Voltage-clamp protocols were applied from holding potentials of −40 to −60 mV. Reversal potentials were calculated from the current-voltage curves. Currents were filtered with a corner frequency of 5 kHz (8-pole Bessel filter), digitized at 20–50 kHz using a Digidata 1320 interface (Axon Instruments, Palo Alto, CA), and acquired in a personal computer using Axoscope software.
Saline solutions for electrophysiology.
We used different saline solutions in the pipette or in the bath, depending on the protocol. Solution A comprised (in mM) 150 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, and 10 HEPES, pH 7.4 (307 mosM); solution B comprised 150 KCl, 5 NaCl, 1 MgCl2, 1 CaCl2, 10 HEPES, and 0.1 EGTA, pH 7.4 (295 mosM); solution C comprised 150 KCl, 5 NaCl, 1 MgCl2, 10 HEPES, and 0.1 EGTA, pH 7.4 (Ca2+-free bath solution; 289 mosM); solution D comprised 132 KCl, 5 NaCl, 1 MgCl2, 15 EGTA (14 mM EGTA-free pipette solution), 1.43 CaCl2 (4.8 nM Ca2+-free pipette solution), and 10 HEPES, pH 7.4 (306 mosM); solution E comprised 139 KCl, 5 NaCl, 1 MgCl2, 8 EGTA (6 mM EGTA-free pipette solution), 1.93 CaCl2 (14 nM Ca2+-free pipette solution), and 10 HEPES, pH 7.4 (314 mosM); and solution F comprised 140 KCl, 5 NaCl, 1 MgCl2, 3.4 EGTA (1 mM EGTA-free pipette solution), 2.33 CaCl2 (100 nM Ca2+-free pipette solution), and 10 HEPES, pH 7.4 (313 mosM). In some experiments, the solution osmolarity was checked with a vapor pressure osmometer (model 5500; Wescor, Logan, UT).
Membrane potential control.
We used solution A inside the pipette and solution B in the bath to record single-channel currents. The rationale for this procedure in the cell-attached configuration was twofold: 1) solution A inside the pipette warranted more physiological ion gradients across the patch so that we could study the ATP-activated phenomena more efficiently, and 2) solution B in the bath completely depolarized the cell membrane (∼0 mV), so it was assured that the real value of the voltage potential across the patch was practically the same as the nominal holding potential applied. This circumvented the inconvenience of having to measure the cell potential to calculate the real holding potential.
In most of the experiments, drugs were administered using an automatic micropipette with variable volume (Gilson, Villiers-le-Bel, France) by positioning its tip inside the bath solution. After the micropipette was in the bath, we waited a few seconds for stabilization of the baseline before drug application. Final concentrations of the agonists or antagonists ranged from 0.001 to 1,000 μM in the bath or in the pipette solution, depending on the protocol. We did some experiments under perfusion conditions (RC-24 chamber; Warner Instrument, Hamden, CT) to confirm the data obtained through micropipette application. All drugs were dissolved in saline solution immediately before use. Dye uptake and ion currents were studied by application of agonists (from 5 to 30 s) just once.
Dye uptake assay.
Cell permeabilization was visualized by the differential uptake of Lucifer yellow (457 Da) and ethidium bromide (394 Da) as previously described (47) or with rhodamine dextran (RITC-Dextran; 11,000 Da). Dyes were dissolved in the bath or in the pipette solution (0.5 mg/ml Lucifer yellow or 0.1 mg/ml ethidium bromide) according to a previously described protocol (47) and applied to the cell for 10 min at ∼37°C. After that, culture dishes were washed three times using normal saline (in mM: 150 KCl, 5 NaCl, 1 MgCl2, 0.1 EGTA, and 10 HEPES, pH 7.4) and observed under a fluorescence microscope (Axiovert 100; Carl Zeiss, Oberkochen, Germany) equipped with rhodamine (Zeiss BP 546/FT 580/LP 590) and fluorescein (Zeiss BP 450–490/FT 510/LP 520) filters. We used Kappa software (Gleichen) for analysis of the images. One or two frames were acquired before agonist application (time 0), and subsequent frames were obtained after cell stimulation. The duration of frame acquisition depended on the kinetics of dye uptake of each cell in each protocol, ranging from 10 to 3,600 s.
Trypan blue exclusion experiments.
Peritoneal macrophages or 2BH4 cells were tested for Trypan blue exclusion during the seal and the protocols. We added ATP, BzATP, and A-23187, thapsigargin, or ionomycin at 37°C. Trypan blue (0.4%) uptake was monitored (45) before the seal and at the end using a light microscope.
ATP, BzATP, cAMP, ADP, UDP, UTP, adenosine, ionomycin, oxidized ATP (adenosine 5′-triphosphate, periodate oxidized sodium salt), apyrase, ethidium bromide, Lucifer yellow, thapsigargin, A-23187, rhodamine B isothiocyate-dextran (RITC-Dextran), pyridoxal phosphate-6-azophenyl-2′,4-disulfonic acid, suramine, and brilliant blue G were purchased from Sigma Chemical (St. Louis, MO). Trypan blue was purchased from Allied Chemical (Detroit, MI), and BAPTA-AM was obtained from Molecular Probes (Eugene, OR).
We normalized fluorescence data for the maximal fluorescence value using Microsoft Excel software and plotted the results using GraphPad Prism software, version 3.0 (San Diego, CA). The calculation of Ca2+ concentration was performed using Sliders version 2.00 and WinMAXC version 2.10 software (Pacific Grove, CA). Data are expressed as means ± SD as indicated. The statistical significance of differences was tested using one-way ANOVA followed by Tukey's test.
Characterization of the pore associated with the P2X7 in 2BH4 cells.
To address whether P2X7 pore formation is dependent on second messengers, we first characterized this receptor in our system. For this purpose, we evaluated 2BH4 cells (a thymic epithelial cell line) and primary culture of mouse peritoneal macrophages that expressed the P2X7 receptor (7, 47). In Fig. 1, we can see that 2BH4 cells and macrophages responded similarly to ATP application, with comparable single currents and dye uptake (Table 1; Fig. 1, C and D). For control, we preincubated 2BH4 cells or macrophages with ethidium bromide or Lucifer yellow without pipette seal (data not shown) or agonist stimulation (Fig. 1, A and B), and no change in fluorescence occurred. After ATP application, 2BH4 cells and macrophages responded with a change in fluorescence for both ethidium bromide and Lucifer yellow (Fig. 1, C and D). The time to reach one-half the fluorescence intensity (t1/2) was not significantly different between these cell types (n = 10, P = 0.063; Table 1). Using the same voltage protocol, we found that 2BH4 cells and peritoneal macrophages showed similar single-current amplitudes and hence similar conductance values (Table 1). In other control experiments with 2BH4 cells, the application of ATP without any dyes in the bath or in the pipette produced single currents with the same features as those found in the presence of dyes (data not shown), suggesting that single-channel currents were not influenced by the presence of dyes (Table 1).
Different ATP analogs were applied to 2BH4 cells.
ATP and BzATP-induced currents and dye uptake (Fig. 1E), whereas other nucleotide analogs did not (data not shown). BzATP was more potent than ATP (EC50 = 600 ± 1.5 nM and 92 ± 7.8 μM, respectively) in inducing pore formation and ion current. This pharmacological profile is consistent with the properties of the P2X7 receptor.
Oxidized ATP (300 μM), described as an irreversible P2X7 receptor blocker (36), was applied to the bath after cell-attached configuration was established. An incubation period of 1 h was allowed, and then 1 mM ATP was also applied to the bath (Fig. 2A). Under such conditions, we expected P2X7 receptors in those cells to be blocked, excluding those located in the patch delimited by the pipette (i.e., not exposed to either oxidized ATP or ATP). As a matter of fact, both single-channel current and dye uptake were blocked. Other blockers such as KN-62, brilliant blue G, and suramin also inhibited dye uptake and ion currents in ATP-stimulated 2BH4 cells (data not shown).
Pore formation associated with the P2X7 receptor requires the action of an intracellular second messenger.
To test for a possible second messenger, 1-h incubation was allowed after establishing the cell-attached configuration with a pipette containing solution A plus 300 μM oxidized ATP plus Lucifer yellow. Subsequently, we applied 1 mM ATP to the bath containing ethidium bromide as illustrated in Fig. 2, B and G (n = 8). Under these conditions, ethidium uptake would be expected, but not Lucifer yellow uptake or ion currents. That would be the case unless the external purinergic stimulus was linked to the membrane patch by means of an internal second messenger, because P2X7 receptors in the patch were blocked. Under this condition, we observed Lucifer yellow and ethidium bromide uptake as well as single-channel currents similar to those in experiments without oxidized ATP (P = 0.019). Figure 2B favors the second messenger hypothesis. If that is true, the other hypothesis might also be possible, and intrapipette ATP at a sufficiently high concentration might be able to drive the production of this second messenger, activating pores around the cell, with concomitant ethidium bromide uptake and recording of whole cell ion currents. Figure 2, C and G, shows that this is the case.
Several authors have demonstrated that mechanical stimulation can release ATP from different cell types (34, 38, 40, 56). Thus the mechanical stimulus provided by the micropipette itself could elicit ATP release outside the membrane patch, promoting P2X7 pore formation. To rule out the possible effects of unnoticed ATP release, we used apyrase (2 U/ml), an ectonucleotidase, in the pipette solution (plus Lucifer yellow) and stimulated the cell membrane outside the patch with 1 mM ATP (Fig. 2D). We also applied apyrase (2 U/ml) in the bath solution (plus ethidium bromide) and stimulated the cell with 1 mM ATP inside the pipette (Fig. 2E). There was no change attributable to apyrase in the single-current recordings or in dye uptake in either condition. The t1/2 of ethidium uptake with apyrase in the bath was similar to the t1/2 of the control with ATP application in the bath without apyrase (139 ± 27 s) (n = 7; P = 0.032), and the t1/2 of ethidium uptake with apyrase inside the pipette was likewise similar to that of control (146 ± 19 s) (n = 6; P = 0.0033). Thus it seems that neither ATP leakage nor ATP released from other sources interfered with our results.
Because P2X7-associated pores are permeable to fluorescent molecules up to 900 Da (52, 53), we evaluated not only ethidium bromide (314 Da) and Lucifer yellow (457 Da) but also propidium iodide (636 Da) and Trypan blue (961 Da) in our permeabilization assays to test for molecular mass. Consistent with activation of the P2X7 receptor, we found that 2BH4 cells treated with ATP were permeabilized to ethidium bromide, Lucifer yellow, and propidium iodide but not to Trypan blue (data not shown). Figure 2F shows that simultaneous use of Lucifer yellow and RITC-dextran (11,000 Da) inside the pipette with ATP stimulus in the bath led to uptake of Lucifer yellow but not RITC-dextran, making clear the selection according to molecular mass and that dye uptake occurred without cell lysis.
Intracellular second messenger characterization.
To identify second messengers involved in P2X7 pore formation, we analyzed the effects of altering the intracellular composition on pore formation using the whole cell configuration. First, we investigated the possible role of Ca2+ as an intracellular signal because data reported in the literature (10, 20, 32) and from our group suggested that Ca2+ might be essential for P2X7 pore formation.
We used different buffered free Ca2+ concentrations in the micropipette solution as shown in Fig. 3. A dose-response relationship between intrapipette Ca2+ and permeabilization became apparent (Fig. 3F). At low Ca2+ concentrations, there was a strong inhibition of dye uptake (n = 10; Fig. 3, D–F). In this situation, we recorded only single currents (n = 9; Fig. 3, D and E), and our calculations indicated that related channels had conductance levels similar to the control values.
As Ca2+ concentration decreased in the pipette solution, currents passed from macroscopic to microscopic and then to single channel. At low Ca2+ concentration in the pipette, there was a large decrease in the permeability of large organic cations such as N-methyl-d-glucamine (NMDG+; data not shown) but not of Na+. Single currents recorded in low pipette Ca2+ concentrations lasted milliseconds (Fig. 3, D and E), in contrast to the ones recorded in 1 mM Ca2+ concentration, with a duration of seconds (Fig. 1, C and D). Accordingly, the t1/2 for ethidium uptake at different intracellular Ca2+ levels was increased as intracellular Ca2+ concentration was reduced.
The rise of intracellular Ca2+ potentiates pore formation.
To further evaluate the need for intracellular Ca2+ for pore formation, we tested whether a Ca2+ ionophore could potentiate the ATP permeabilization effect by associating ATP in low, ineffective concentrations (when alone) in inducing dye uptake or ion currents plus low, ineffective concentrations of ionomycin. Even though 25 μM ATP alone (n = 8; Fig. 4, A and G) or 1 μM ionomycin alone (n = 6; Fig. 4, B and G) were ineffective, when applied together, they triggered dye uptake and single currents with kinetics similar to those of 100 μM ATP (t1/2 = 160 ± 18 s) (n = 6; Fig. 4, C and G). Accordingly, 15 μM ATP in association with 1 μM ionomycin-triggered dye uptake with kinetics similar to those of 50 μM ATP (177 ± 27 s) (n = 6; data not shown).
The specificity of such an association for the P2X7 receptor pathway was assessed by preincubating 2BH4 cells with 300 μM oxidized ATP for 1 h (n = 5; Fig. 4D) or 10 μM brilliant blue G for 10–15 min (n = 4; data not shown) before applying ATP plus ionomycin. As a result, both ethidium uptake and single currents were inhibited (Fig. 4, D and G).
When used alone, ionomycin induced pore formation in a dose-response pattern, with t1/2 = 118 ± 17 s (n = 11; P = 0.022), which is similar to that obtained with ATP alone (data not shown). Consistently, 10 μM ionomycin alone induced ion current amplitudes similar to those of 1 mM ATP alone (−31 ± 6 pA, n = 11, and −29 ± 5 pA, n = 12, respectively; compare Fig. 4C and Table 1). The same is true for conductance (412 ± 204 pS, n = 5; data not shown). As expected for ionomycin, pore activation was abolished if a Ca2+-free solution was used in the bath (Fig. 4F). Trypan blue exclusion was used to assess viability throughout the experiments described above.
Other molecules that act on Ca2+ homeostasis can give rise to another pore with some of the characteristics of the P2X7-associated pore.
A-23187, another Ca2+ ionophore, also induced dye uptake with kinetics consistent with the pore associated with P2X7 activation (Fig. 5A). However, A-23187 was less potent than ionomycin as an inducer of pore formation. In addition, A-23187 exhibited slower kinetics of ethidium uptake (A-23187 t1/2 = 188 ± 41 s, n = 7) than that observed for ATP (t1/2 = 125 ± 15 s, n = 12; P = 0.00022) or ionomycin (t1/2 = 118 ± 17 s, n = 11; P = 0.00065).
Thapsigargin, which specifically blocks endoplasmic reticulum Ca2+ ATPase (depleting intracellular stores and hence raising cytosolic Ca2+ concentration) also resulted in dye uptake similar to that obtained with ATP (Fig. 5B). The kinetics of ethidium uptake induced by thapsigargin was slower than those of ATP or ionomycin (thapsigargin t1/2 = 184 ± 19 s, n = 5). However, thapsigargin-induced dye uptake was maximal after 6 min, which is similar to ATP application.
In our experiments, A-23187- and thapsigargin-induced single currents exhibited conductance similar to that obtained with ATP or ionomycin application (Fig. 5, A and B). Consistently, BAPTA-AM (a Ca2+ chelator permeable through membrane) was able to inhibit both dye uptake and single currents when used in combination with ATP (Fig. 5C). BAPTA-AM also inhibited pore formation induced by ionomycin, A-23187, and thapsigargin in the thymic epithelial cell line IT45-R1 and in human embryonic kidney (HEK) cells (data not shown).
In view of the latter results, we treated IT45-R1 and HEK cells, which do not express P2X7 receptors, with ionomycin, A-23187, and thapsigargin, resulting in dye uptake similar to that observed in 2BH4 cells (data not shown). In addition, 300 μM oxidized ATP did not inhibit such dye uptake. Taken together, these results suggest that intracellular Ca2+ may activate another pore not related to the P2X7 pore, similarly to the pore activated or formed by maitotoxin as described elsewhere (35).
MAP kinases are essential for pore formation.
Several groups have shown that P2X7 receptor activation can signal through MAP and MEK kinases in diverse cell types (4, 9, 10, 19, 20, 49). For this reason, we tested whether MAP kinases play a role in the pore formation of the P2X7 receptor.
We pretreated 2BH4 cells with an inhibitor of MAPK and MEK for 15 min before addition of 1 mM ATP. SB-203580, a blocker of p38 MAPK, potently inhibited ATP-induced dye uptake in 2BH4 cells and peritoneal macrophages (Fig. 6) with IC50 values of 75 ± 2 nM (n = 7) and 104 ± 5 nM (n = 4), respectively (Fig. 6D). PD-98059, a MKK inhibitor (39), also potently blocked ATP-induced pore formation in 2BH4 cells (Fig. 6C) and peritoneal macrophages (Fig. 6E), with IC50 of 88 ± 3 nM (n = 8) and 123 ± 17 nM (n = 4), respectively (Fig. 6E). MAPK antagonists also blocked single currents in 2BH4 cells (Fig. 6, B and C) and macrophages (data not shown). Our positive control with 1 mM ATP, but without MAPK antagonists, is shown in Fig. 6A. Taken together, these data strongly suggest that MAPK and MKK are vital for pore formation associated with the P2X7 receptor both in 2BH4 cells and macrophages.
Is the pore associated with P2X7 receptor a gap junction hemichannel?
Gap junction hemichannels are permeable to molecules up to 1,000 Da (6). So, we ruled out the possibility of the participation of hemichannels using hemichannel inhibitors such as heptanol, carbenoxolone, and 18α-glycyrrhetinic acids (11, 15, 29). None of these hemichannel inhibitors blocked the pore formation induced by ATP (Fig. 7), ionomycin, or thapsigargin (data not shown). As investigators at our laboratory (2) previously demonstrated, gap junction hemichannels are not part of the P2X7 receptor permeabilization response.
The mechanism of ATP-induced pore formation has remained unresolved since its first report in mast cells (5a). At least two possibilities have been raised in the literature. The first is based on the idea of dilation of a channel that would incorporate subunits, thus producing a growing pore that would allow the passage of progressively larger (∼900 Da) cations (3, 14, 31, 46, 52, 54). The second is based on the independent existence of a channel permeable to small cations (16) and a pore permeable to larger cations and fluorescent dyes (8, 17, 22, 41, 44). In the latter situation, a means of signaling would be necessary to link the ATP receptor (i.e., channel permeable to small cations) to the pore. A possible signaling pathway could involve the existence of second messengers. In our study, we aimed to investigate precisely the role of second messengers on such pore formation.
First, we have demonstrated that the P2X7 receptor was responsible for pore formation in 2BH4 cells and peritoneal macrophages as previously described (7, 14). Ion current recordings (single currents) were obtained after ATP stimulation. These currents showed values of conductance (445 ± 8 pS) consistent with data described previously by Coutinho-Silva and Persechini (17) in murine macrophages (409 ± 33 pS). The kinetics of fluorescent dye uptake obtained in our experiments (maximal response in 2–5 min) had a latency of 20–40 s, which is consistent with previous reports (26, 28, 31). The t1/2 of dye uptake was 129 ± 16 s in 2BH4 cells and 127 ± 11 s in peritoneal macrophages, also similar to previous data (24, 53, 54). We showed that this ATP- or BzATP-induced pore formation was blocked by preincubation with oxidized ATP and other blockers (data not shown) in the bath as expected for a P2X7-triggered phenomenon.
The pore associated with P2X7 depends on second messengers.
The cell attached configuration was an important tool in our work in assessing the role of second messengers because of its unique features: 1) it allows for controlled compartmentalization of the intra- and extracellular milieux, and 2) it isolates membrane patches, creating specific areas that can be surrounded by other areas exposed to different pharmacological stimuli.
Taking advantage of the cell-attached condition, we observed that pore formation occurred both in the patch and in the rest of the membrane, despite preincubation of cells with oxidized ATP in the pipette solution at a concentration previously shown to block P2X7 receptors (Fig. 2B). The response was the same when oxidized ATP was used in the bath and ATP applied in the pipette solution (Fig. 2G). By using apyrase (an ectonucleotidase) in similar conditions (Fig. 2, D and E), we ruled out the idea that this effect could be a mere artifact due to mechanical stimulation, a well-known mechanism of ATP release in most cell types.
Thus our data strongly suggest that pore formation occurred by the action of an intracellular second messenger. These results are in agreement with those found in a human macrophage cell line (THP-1) in which a different methodology was used, clearly demonstrating that second messengers were necessary to form the P2X7-associated pore (19), discussed below.
Because the P2X7 receptor is capable of triggering a long-lasting transmembrane Ca2+ influx, we investigated whether Ca2+ could be the second messenger involved in the pore formation. For this purpose, the whole cell configuration was used instead because of the possibility of strict control of the intracellular milieu. Several different Ca2+ concentrations were used in the pipette solution, making clear that a decrease in the intracellular Ca2+ concentration caused inhibition of the pore formation in a concentration-dependent manner (Fig. 3). Furthermore, we observed that there was an increase in t1/2 value (t1/2 = 178 ± 36 s) and longer latencies in the fluorescent dye uptake.
These findings suggest that Ca2+ may be an important cytoplasmic factor involved in the coupling of channel activation and pore formation. Other groups previously had indicated the participation of Ca2+ in the intracellular signaling of P2X7 receptor (43, 44, 12, 19), but no direct measurement of Ca2+ dependence for pore formation had been published previously. In contrast, Virginio et al. (52), working with rat P2X7, did not find dependence of intracellular Ca2+ when they withdrew intracellular Ca2+ in a whole cell configuration.
The hypothesis of intracellular Ca2+ participation was reinforced when we pretreated cells with BAPTA-AM before application of ATP and BzATP. In this case, pore formation was strongly inhibited, regardless of the agonist applied. Our results clearly indicate that Ca2+ is necessary for pore formation in 2BH4 cells and macrophages (data not shown for the latter). On the other hand, Virginio et al. (53) showed that HEK cells expressing P2X2 receptors did not show a significant change in NMDG+ permeability after incubation with either BAPTA-AM or BAPTA in the pipette. This suggests that the pore associated with different subtypes of P2X receptors may be regulated by different mechanisms.
In keeping with the idea that Ca2+ is a second messenger, we have shown a synergic effect between ATP and drugs that raise intracellular Ca2+ (Fig. 4). Because Ca2+ signaling is generally complex, this result may be due to a phenomenon that does not depend solely on Ca2+ concentration in a given time frame, but instead depends on the frequency of Ca2+ increase and on the subcellular localization of such increase. More recently, Lajdova et al. (32) showed that the rise of intracellular Ca2+ triggered by 4-aminopyridine was mediated by P2X7 receptor, because this effect was inhibited by P2X7 receptor blockers.
MAP kinases can participate in the process of pore formation.
Some groups have reported the activation of MAP kinases after stimulation of the P2X family of receptors (27, 50). More recently, Gendron et al. (20) showed that activation of the P2X7 receptor could lead to phosphorylation of extracellular signal-regulated kinases (ERK1/2) that were blocked by Ca2+ chelators when applied in an intra- or extracellular milieu. Budagian et al. (10) found similar results by using EGTA to study the activation of p56lck or MAP kinases mediated by the P2X7 receptor. Gudipaty et al. (23) showed that interleukin-1β secretion is dependent on influx of extracellular Ca2+ and a sustained rise in cytosolic Ca2+ in monocytes, macrophages, and HEK-293 cells. These studies did not investigate the relationship of MAP kinases to the pore associated with P2X7. In the present study, we tested whether MAP kinases play a role in the pore formation of P2X7 receptor. This is not a detailed study; however, our data using pharmacological tools suggest that MAP kinases are essential to the process of pore formation. These results are in agreement with those of Donnelly-Roberts et al. (19).
There are other pores besides the P2X7 pore in the membrane.
There is evidence in the literature that other molecules similar to P2X7 receptors can form and/or induce pores in cell membranes, but they belong to different classes of proteins. For example, connexins (the proteins that form gap junctions) can form hemichannels that permeabilize several cell types (6), and maitotoxin is associated with another membrane pore that can be distinguished pharmacologically from the P2X7-associated pore (35). The maitotoxin-associated pore depends on an increase in intracellular Ca2+. Similarly, we have demonstrated that drugs such as ionomycin, A-23187, or thapsigargin that raise intracellular Ca2+ induce pore formation. Therefore, these data suggest that drugs that increase intracellular Ca2+ may activate a membrane pore. Our results suggest that these pores are different from the P2X7-associated pore because they are not inhibited by P2X7 blockers.
In summary, we have demonstrated for the first time, by using a combination of electrophysiological and dye uptake techniques, that second messengers (Ca2+ and MAP kinases) are necessary for the pore formation associated with the P2X7 receptor. Thus we have opened new avenues for the study of the mechanisms of pore formation associated with P2X7 receptor activation.
We thank Dr. Wilson Savino for the use of his laboratory's facilities and Drs. Robson Coutinho Silva, David C. Spray, Rodrigo C. Bisaggio, and Oscar Kenji Nihei for critical review of the manuscript and helpful discussions.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society